Download Friend or foe? Crowdfunding versus credit when banks

Friend or foe? Crowdfunding versus credit
when banks are stressed
Daniel Blaseg a and Michael Koetter b,c,∗
a Goethe-University Frankfurt, Theodor-W.-Adorno-Platz 4, 60323 Frankfurt am Main, Germany
b Frankfurt School of Finance and Management, Sonnemannstr. 9-11, 60314 Frankfurt a.M., Germany
c Halle Institute for Economic Research, Kleine Maerkerstrasse 8, 06108 Halle (Saale), Germany
This version: 15 April 2016
Abstract
Bank instability seemingly could push borrowers to use crowdfunding as a source
of external finance. A range of stress indicators help identify banks with potential
credit supply constraints, which then can be linked to a unique, manually constructed sample of 157 new ventures seeking equity crowdfunding, for comparison
with 200 ventures that do not use crowdfunding. The sample comprises projects
from all major German equity crowdfunding platforms since 2011, augmented with
controls for venture, manager, and bank characteristics. Crowdfunding is significantly more likely for new ventures that interact with stressed banks. Innovative
funding sources are thus particularly relevant in times of stress among conventional
financiers. But crowdfunded ventures are generally also more opaque and risky than
new ventures that do not use crowdfunding.
Keywords: equity crowdfunding; credit crunch; bank stress
JEL classification: G01; G21; G30
? We are grateful for feedback received from the Southern Finance Association
Annual Meeting and the P2P Financial Systems Bundesbank/SAFE/UCL conference and seminars at Goethe University. Comments from our discussants Karsten
Wenzlaff and Marc William Simpson and other seminar participants are gratefully
acknowledged. We thank in particular Christoph Siemroth, Bernd Skiera, and Adalbert Winkler for their detailed feedback. All errors are our own.
∗ Corresponding author. Tel.: +49 69 154 008 446.
E-mail addresses: [email protected] (D. Blaseg), [email protected] (M.
Koetter).
1
Introduction
Akerlof’s (1970) seminal lemons problem epitomizes the key challenge faced
by any investor: how to select projects from a pool of opaque applicants. Traditionally, banks help resolve the information asymmetry between savers and
investors by developing screening competences and acting as delegated monitors (Diamond, 1984). But dramatically reduced transaction and information
acquisition costs, together with historically low interest rates, impede banks’
incentives to engage in costly information generation, which can lead to the
contraction of credit (Puri et al., 2011; Jiménez et al., 2012) or misallocated
funding to too risky projects (Dell’Ariccia and Marquez, 2004; Jiménez et al.,
2014). Against this backdrop, recent studies by Belleflamme et al. (2013) and
Mollick (2014) hypothesize that crowdfunding may rival bank finance and connect even small savers with risky new ventures that face traditionally tighter
financing constraints (e.g., Cassar, 2004; Robb and Robinson, 2014).
In turn, we test whether the wisdom-of-the-(investor)crowd can substitute
for bank credit as a major source of funding for new ventures, by exploiting exogenous shocks to young ventures’ banks. We construct a novel, hand-collected
data set of ventures’ uses of equity crowdfunding in Germany, their relationships with banks, and various venture traits since 2011. By observing venturebank relationships, we can identify if ventures connected to shocked banks are
more likely to use crowdfunding in an attempt to substitute for contracting
bank credit supply. In so doing, we move beyond the admittedly important
descriptive evidence in this nascent strand of literature, which does not permit
inferences about the causal effects of the determinants of crowdfunding. 1
1
Recent policy (e.g., De Buysere et al., 2014), and academic (e.g., Mollick, 2014;
2
We also control for observable management and venture traits to determine
if more opaque ventures with greater information asymmetries are more likely
to use crowdfunding as an alternative source of financing. Greater information
asymmetries increase capital costs, which implies a well-known pecking order
of capital structure: Internal funds are preferred over debt, and equity is a last
resort of funding (Jensen and Meckling, 1976; Myers and Majluf, 1984). To
mitigate information asymmetries and facilitate the efficient allocation of financial resources, from savers to productive investors, financial intermediaries
can generate private information by establishing close and long-term relationships (Rajan, 1992; Uchida et al., 2012). But relationship lending is costly,
so banks may turn down funding requests by promising, yet hard-to-assess
projects such as new ventures if they cannot confidently cover the costs associated with producing necessary private information (Rajan, 1992; Petersen
and Rajan, 1994, 2002). In this setting, we investigate if ventures tied to
banks that struggle to cover the costs of private information generation are
more likely to tap a potentially less-than-wise crowd as a funding source.
The financial crisis of 2008 amplified the generally prevalent challenges that
young and small ventures confront when trying to raise external finance. In the
aftermath of the great financial crisis, the number and volume of equity financing rounds from venture capital sources declined significantly (Block et al.,
2010), credit supply tightened in the Eurozone (Hempell and Kok, 2010),
and in Germany, even local lenders reduced their loans (Puri et al., 2011).
Gorman and Sahlman (1989) and Cassar (2004) caution that credit supply
Schwienbacher, 2013; Hornuf and Schwienbacher, 2014) light on the potential role
of crowdfunding and vividly illustrate the broadening interest in this new form of
financing ventures. We instead seek to provide empirical evidence about the causal
effects of bank credit crunches.
3
shocks are especially important for new ventures. However, most existing empirical evidence is geared toward venture capitalist funding (for an overview,
see Gompers and Lerner, 2001). The ability of crowdfunding to substitute for
bank credit or other sources of external finance, due to its significantly lower
transaction costs in the Internet age, in particular remains unclear.
This research gap exists primarily because of the absence of data. We handcollected a sample of all the ventures that applied for funds on major German
equity crowdfunding platforms since 2011. That is, among 357 new ventures
for which we have data, 157 applied for equity crowdfunding at one of the six
major German online platforms between November 2011 and June 2014, which
cover 95% of the total market in terms of offerings and 99% in terms of volume.
Figure 1 illustrates the structure of the sample and the main specifications that
explain the odds that a venture apply for external funding on a crowdfunding
platform conditional on its bank relationship and venture and management
traits.
– Figure 1 around here –
We manually gathered the data for the crowdfunding ventures from each platform webpage and database. For the 200 ventures that did not use crowdfunding, we obtained the venture and management variables from the membership
database of the Federal Association of Startups. Thus, in contrast with previous research into crowdfunding (e.g., Belleflamme et al., 2013; Mollick, 2014),
we can estimate the probability of tapping the “wisdom of the crowd”, conditional on venture and managerial traits, relative to a relevant comparison
group of comparable young ventures that face similar financing constraints.
Another challenge that plagues empirical literature pertaining to the role of
4
crowdfunding is the notorious unobservability of the arguably most important
competing source of external finance: bank credit. Because we collect information about each ventures’ bank relationship, we can exploit the heterogeneity
in bank distress in the aftermath of the financial crisis and identify credit
supply shocks to ventures, according to the health of their main external financier. To our knowledge, this article is the first to seek to identify the effect
of bank stress on alternative forms of external finance directly.
In total, we identify 82 banks connected to the new ventures in our sample and specify five alternative indicators of stressed relationship lenders. The
main indicator is whether a bank received capital support from the German
Special Fund for Financial Market Stabilization (“SoFFin”), which came into
effect as of 2008. With an alternative approach, we also classify banks as
stressed if they report an existing restructuring plan, according to the comprehensive assessment conducted by the European Banking Authority (EBA)
in November 2014, how stable the bank is according to the EBA, how stable
the parent of a bank is according to the EBA, and whether a regional savings
bank belongs to a stressed Landesbank in 2008 (see Puri et al., 2011).
The main results show that ties to a bank bailed out by the SoFFin increase
the probability that the venture taps a crowdfunding platforms by 18%. The
probability of successfully completing a crowdfunding request increases by
22% tough, so the successful completion of a crowdfunding request (the left
branch in Figure 1) does not appear to depend on the indicators of bank
distress. That is, credit supply shocks determine the choice to seek alternative
funding forms, but they do not necessarily discriminate between projects that
can or cannot convince the crowd. The positive effect of crunched banks the
use of crowdfunding remains statistically and economically significant, even
5
when we control directly for bank financial profiles. Alternative indicators of
bank distress, and especially the existence of restructuring plans shared with
the EBA, yield qualitatively similar results, though with weaker statistical
significance. Regarding other venture and management traits, we find that the
likelihood of using crowdfunding is significantly larger for ventures that exhibit
lower ratings, are smaller, and have fewer tangible assets. This result may
indicate that ventures with greater information asymmetry suffer the most
from a credit supply shock, and therefore seek crowdfunding as an alternative.
Whether these projects are more likely to be lemons or gems that have been
neglected by banks is an important question for further research.
The remainder of this article is organized as follows: Section 2 relates our
study to prior literature and provides an institutional background of equity
crowdfunding in Germany. In Section 3, we present and discuss crowdfunding
data, as well as our identification strategy for bank-venture relationships. We
discuss the empirical findings in Section 4 and conclude in Section 5.
2
Literature and background
2.1
Bank funding and crowdfunding
Banks are vital to resolve information asymmetries, especially those that
plague small and medium enterprises (e.g., Petersen and Rajan, 1994, 2002;
Berger and Udell, 1998). The quality of opaque new ventures is difficult for
investors to evaluate and information asymmetries always exist during external, early stage financing (see Jensen and Meckling, 1976; Stiglitz and Weiss,
1981; de Meza and Webb, 1987). Information asymmetries between ventures
6
and possible investors result in the well-known pecking order of capital (Myers
and Majluf, 1984), such that ventures prefer to finance new projects with retained earnings or other internal cash flows, because external funds are more
expensive. External debt financing is favored over equity, because the latter
dilutes the ownership of the entrepreneur. Robb and Robinson (2014) use the
Kauffman Firm Surveys to document the important role of debt at the beginning of a venture’s life and suggest that the largest part of total capital comes
from outside debt, followed by owners’ equity, then insider debt, outside equity, and finally owner debt. Brown et al. (2012) also note the important role
for bank debt as a source of funding for new ventures in Germany.
The financial crisis aggravated the financing challenges faced by young ventures during and after 2008 (e.g., Popov and Udell, 2012; Jiménez et al.,
2012). Puri et al. (2011) document a credit supply crunch among German
local lenders and Hempell and Kok (2010) identify a significant bank lending
contraction in Germany from the ECB lending survey. Considering the important role of debt use in entrepreneurial financing, we conjecture that banks
transmitting a credit shock may cause the young ventures connected to them
to grow more inclined to find new sources of funding, especially if small financing volumes imply high relative transaction costs that are unattractive to
large-scale investors (Titman and Wessels, 1988; Robb and Robinson, 2014).
A novel way to reduce transaction costs in entrepreneurial financing is
crowdfunding. Schwienbacher and Larralde (2010) provide an overview of
nascent equity crowdfunding literature in relation to entrepreneurial finance,
in which they discuss why founders choose this source of funding. Hornuf and
Schwienbacher (2014) and Mollick (2013) compare crowdfunding to different
entrepreneurial financing options. Hemer (2011) emphasizes that the funding
7
process itself is the decisive difference, because ”entrepreneurs make an open
call for funding on a crowdfunding platform, and investors make their decisions based on the information provided therein. Moreover, the crowdfunding
platform facilitates the transaction by providing a standardized investment
contract and settling the payments.” Bradford (2012) defines equity crowdfunding as a scenario in which supporters or investors receive a stake in the
ventures they fund, in the form of profit participation or straight equity. We
similarly define equity crowdfunding as a source of funds , obtained when an
entrepreneur sells equity shares of a company to a group of (small) investors
through an open call for funding on Internet-based platforms.
2.2
Institutional background
Equity crowdfunding platforms are non-bank financial institutions that
provide intermediation services for the offering and sale of stocks and similar securities to the general public. These services include the provision of
standardized contracts, technology infrastructure for the transactions, and investor relations. To reduce investors’ transaction costs, they also provide standardized information, such as pitch decks, financials, and valuations sourced
from the venture, without guaranteeing their correctness though. Most equity
crowdfunding platforms do not act as open marketplaces but instead serve as
network orchestrators, curating the offerings placed on the platform after a
cross-check of formal criteria, such as limited liability and available documentation.
Whereas some platforms allow the direct acquisition of securities in the
venture, others act as nominated agents and pool funds. Because they facilitate the sale of equity-like instruments without voting rights, the platforms
8
fall outside legal brokerage framework, though rapidly growing crowdfunding markets worldwide have prompted some countries (e.g., Italy, the United
Kingdom, France, Germany, Spain) to develop specific crowdfunding regulations, with the goal of protecting unprofessional investors and increasing the
transparency of offers in the shadow banking market.
German crowdfunding platforms use financial instruments and equity-like
mezzanine capital, such as silent partnerships (Stille Beteiligungen) and participation rights (Genussrechte). More common debt-like mezzanine instruments take the form of subordinated loans (Partiarische Nachrangdarlehen),
which are less regulated. The offerings of a venture based on equity-like securities in Germany are limited to EUR 100,000 per year without an official
prospectus, which is accepted by the Bundesanstalt fuer Finanzdienstleistungsaufsicht (BaFin) as long as there are more than 20 investors or the
offering is aimed at unprofessional investors with a share price of less than
EUR 50,000. Subordinated loans skirt this problem and allow offerings with
higher volumes.
As an intermediary between investors and the ventures looking for funding,
the platforms are not directly involved in the financial activity and take on very
limited responsibility. Revenue is mostly generated from the success fees for
offerings that exceed their minimum requested amount, which range between
5% and 10% of the amount raised. Few platforms operate as full banks, which
means they cannot handle the payments on their own and instead must engage
an authorized payment service provider or bank, which incurs additional costs
of 1% - 3% for the funded venture. Expenses to produce a video, often a core
element in an offering, together with the costs of preparing and running the
campaign and maintaining the investor relations afterwards, also must have
9
to be taken into account by the venture.
– Table 1 around here –
Table 1 provides an overview of the German crowdfunding market. The first
six projects were funded at the end of November 2011 on the Innovestment and
Seedmatch platforms. As of December 2014, 14 active crowdfunding platforms
were facilitating equity crowdfunding or revenue-sharing models in Germany.
Nine more platforms started operations but closed before their first offering.
The total funding volume of equity crowdfunding platforms in Germany in
2011 was around EUR 0.45 million, but it rose to EUR 35.3 million by the
end of 2014. Seven of the 14 active platforms had one or no offerings during this
period, and 95% of the total volume was raised on five platforms: Seedmatch
(approximately EUR 19 million), Innovestment (EUR 2.3 million), Bergfuerst
(EUR 4.1 million), Fundsters (EUR 1 million), and Companisto (EUR 7.1
million). In total, 171 offerings by the end of 2014 came from 165 different
ventures. Thirteen offerings were unsuccessful in that the minimum amount
the venture requested by the company was not raised during the funding
process.
3
Sampling and identification
3.1
Sampling
To identify the differential effect of a credit supply shock on the inclination
of ventures to seek crowdfunding, we sample new ventures that use or that do
not use crowdfunding, as shown in Figure 1. We begin with the members of the
Federal Association of Startups in Germany (Bundesverband Deutsche Star10
tups). It had 264 members by the end of 2014, of which 64 used crowdfunding.
The formal prerequisites to be listed on a German crowdfunding platform are
very similar to those required for a membership in the association. We thus
identified 93 crowdfunding offerings with available information that applied
for funding through the German crowdfunding platforms Bankless24, Bergfuerst, Companisto, Fundsters, Innovestment, Mashup Finance, or Seedmatch
between November 2011 and December 2014. The resulting sample included
157 ventures that used crowdfunding (Group 1) and 200 ventures that did
not (Group 2). Figure 1 also indicates, which of these ventures completed the
funding request. The comparison of their descriptive statistics confirms that
we compare very similar ventures.
We obtain the data by continuously pulling information from each crowdfunding platform’s webpage. The dependent variable is an indicator equal to
“1” if the venture attempted to obtain external finance through crowdfunding
and “0” otherwise. Of all 157 offerings, 85% were successful, and the ventures
raised about EUR 200,000 from 280 investors on average. With an average
company valuation of approximately EUR 1.95m, the investors acquired about
10% of a venture. Before turning to the ventures’ characteristics, we explain
how we collected the data about the venture-bank relationships that we used
to identify the effects of bank stress on the odds of using crowdfunding.
3.2
Identification through bank bailouts
To assess the role of equity crowdfunding as a way to mitigate credit constraints of young ventures, we seek to compare the conditional likelihood that
new ventures seek crowdfunding, according to whether they are tied to healthy
banks or stressed banks.
11
To this end, we collect bank-venture relationships for all 357 ventures from
the Creditreform database. For each company, it provides a unique bank identification number that indicates the financial institutions with which it has a
major credit relationship. We combine these data with the BaFin database
to control for consolidation and obtain complete bank names. In total, we
identify 82 banks (see Table 10), which we categorize as stressed or healthy,
according to the five alternative criteria illustrated in Figure 2.
– Figure 2 around here –
For the 82 banks, our base-line identification defines stressed banks as those
that received equity support from the SoFFin. In October 2008, the German
Federal government founded the Special Fund for Financial Market Stabilization (SoFFin) in response to the turmoil in the aftermath of the collapse of
Lehman Brothers. The fund was designed to strengthen the capital base of
German banks that were hit by taking over problematic positions and providing other guarantees. It had supported a total of 10 German banks since
its inception, with a total volume of outstanding equity and guarantees of
192 billion Euros in 2009. By the end of 2014, the SoFFin remained exposed
to three German banks with share and hybrid capital equivalent to a total
volume of about EUR 17bn.
We matched the bank names from the Creditreform database with public
information about which banks were supported by the SoFFin. However, ventures may self-select into bank relationships depending on the health of that
bank. For example, participating in the SoFFin support program may induce
certain entrepreneurs to avoid seeking credit from such a bank.
Therefore, we also define stressed banks by using the comprehensive assess12
ment by the European Banking Authority (EBA) that took place in November
2014, which is after we observe the crowdfunding choices of new ventures in
this sample. The assessment by the EBA cannot by itself indicate credit supply strain; rather, it offers a testimony of systemic relevance. Our first EBA
based measure of stress is therefore whether a bank had a restructuring plan in
place before 2013. As illustrated in Figure 2, we distinguish generally between
banks assessed directly by the EBA and those connected to a bank holding
company that was assessed. Finally, we consider any regional savings banks
connected to a Landesbank distressed, because their responsible bank holding
company was exposed to the US subprime market shock (Puri et al., 2011).
We also follow Berger and Udell (2004) and calculate CAMEL (i.e., capital,
asset quality, management quality, and liquidity) covariates for every bank,
which we use as proxies for its financial health. Table 2 offers an overview of
the descriptive statistics of the CAMEL covariates, separated by the different
stress indicators.
– Table 2 around here –
Banks that are supported by the SoFFin or have an affiliation with a
stressed Landesbank exhibit worse CAMEL profiles than non-supported or unaffiliated banks. Table 3 summarizes the results of the EBA’s EU-wide stress
tests as additional indicators of stressed banks. Of the 82 banks in the sample, 6 were assessed directly. The parents of another 36 banks were assessed
indirectly. Half of the directly assessed banks had a restructuring plan before
2013, and their Common Equity Tier 1 ratio in the fully loaded adverse scenario was 7.1% on average. The indirectly tested banks also exihibited similar
traits.
13
– Table 3 around here –
3.3
Venture and crowdfunding traits
Table 4 provides an overview of the crowdfunding offerings of the ventures
in the sample that had no missing values. Horizontally, we distinguish three
panels. The first depicts the traits of firms that use and do not use crowdfunding, such as firm size and other factors motivated by venture capital literature
and discussed more extensively in the Results section. Within each panel, we
depict the descriptive statistics for ventures with a relationship to a bank that
is supported by the SoFFin, which is our main indicator of bank distress.
– Table 4 around here –
Regarding venture characteristics, we find that crowdfunding users with ties
to stressed bankers tend to exhibit higher asset tangibility, are significantly
less often located in cities, and have better credit ratings. Yet the ventures do
not differ in terms of size, female board participation, the number of board
members, or receipt of a supporting scholarship from the federal government.
The right-hand panel also clearly illustrates that none of the differences between firms tied to stressed versus healthy banks are significantly different
when we compare crowdfunders with non-crowdfunders. Thus, we need a statistical approach to identify the factors that predict which type of ventures
use crowdfunding.
The second panel shows the crowdfunding characteristics. New ventures
did not differ significantly in terms of crowdfunding volumes, the number of
investors, or firm valuations in the comparison of firms tied to stressed banks
versus those connected to non-SoFFin banks. The only significant difference
14
is the lower success rate of obtaining the aimed volume when the bank of
a venture is stressed. Finally, the third panel shows that our indicator of
bank government support in 2008, SoFFin, effectively gauges the significant
difference in financial profiles, reflected by the so-called CAMEL profiles of
banks. We discuss the individual effects of these variables subsequently; here,
we limit ourselves to noting the upshot of this result: Banks supported by the
SoFFin differ significantly, and these differences should help predict, which
firms use crowdfunding as a substitute for bank finance.
4
Model and results
4.1
Specification and baseline results
Table 5 contains the descriptive statistics for our main test variable, an
indicator variable (soffin) that takes a value of “1” if the bank is supported and
“0” otherwise. In total, 24% of all ventures in the sample have a relationship
to a bank supported by the SoFFin. However, the share of companies whose
bank is supported by the SoFFin is 37% among the group of ventures that used
crowdfunding – more than twice the share of the group of ventures that did
not use crowdfunding (15%). A venture facing larger credit constraints thus
appeared more likely to apply for crowdfunding, after we control for several
venture traits, as we discuss shortly.
– Table 5 around here –
We predict the likelihood that a venture i applies successfully for crowdfunding yi = 1, conditional on venture traits xi and whether it is tied to a
bank that was bailed out by the sof f ini . We use a logit model as a baseline
15
specification and estimate:
Pr(y = 1 | x) =
exp (α + βx)
1 + exp (α + βx)
(1)
In addition to our main variable to test for SoFFin support, we added
the covariates described in Table 11 and summarized in Table 4 step-by-step.
Table 6 contains the marginal effects of the baseline logit regression model to
explain crowdfunding.
– Table 6 around here –
A range of goodness-of-fit indicators, the Pseudo R2 , and Nagelkerke’s R2
support the good discriminatory power of the model, despite the relatively
low sample size (Hosmer and Lemeshow, 2012). We also compare the predicted probabilities against a moving average of the proportion of cases using
a locally weighted scatterplot smoothing graph, which confirms the fit of the
model. Likewise, the area under the receiver operating characteristic curve
(”AURROC”) of 0.84 for Column 7 in Table 6 strongly indicates that the
probability of using crowdfunding is explained quite well by the covariates.
The comparison of the coefficients across ordinary least squares, logit, and
probit models tells a qualitatively similar story about the impact of a regressor
on the probability of crowdfunding. Robust estimation procedures are qualitatively similar, mitigating potential misspecification concerns. Henceforth, we
report the results from the logit regressions.
The marginal effect of the main variable of interest shows that the likelihood
of applying for crowdfunding increases when a venture’s bank is supported by
the SoFFin. The marginal effect is positive and statistically significant in all
16
models. Economically, the effect in column (7) is also important. If a new
venture is connected to a bank supported by the SoFFin, the probability that
it applies for crowdfunding increases by 17.5%. Against the backdrop of an
unconditional probability to apply for funds of 43.9% (=157/357), this effect
is large.
These results support the hypothesis that young ventures are more likely to
tap innovative, alternative sources of external funding, especially then when
their conventional providers of credit are stressed. To assess whether this result
is driven by observable traits related to the degree of information asymmetries
and the quality of the venture, we discuss individual control variables next.
4.1.1
Credit Scores
Credit scores are a common tool that banks use to evaluate ventures’ loan
applications, but it is unclear if these ratings affect the availability of debt
for young ventures. Robb and Robinson (2014) explore this question with
U.S. data from the Kauffman Survey, and observe that information about the
ventures’ past payment behavior can have a negative effect on access to finance
among young ventures. Brown et al. (2012) confirm this view and suggest that
information provided by an external credit agency can affect the availability of
financing for young ventures; ventures with a good rating have better chances
of obtaining a loan, whereas ventures with bad ratings face difficulties getting
a loan. In line with prior literature, we expect that ventures with bad credit
scores are more likely to use crowdfunding.
External credit ratings provided by Buergel range from A (good) to C (bad).
The underlying variable (credit) is coded accordingly, such that rating class A
takes a value of ”1”, indicating that the business relation is approved; rating
17
class B is coded ”2”, which covers approvable business relations and class C is
coded with ”3”, or a bad rating, which means that the business relation is a
matter of trust or discretion. Buergel is one of the largest databases on German
companies, with more than 3.9 million entries. With BoniCheck, a product of
Euler Hermes, it offers an instrument for assessing ventures’ solvency. From
the Buergel database, we deduced whether an external credit rating, in the
form of the BoniCheck indicator, was provided for each venture and, if so,
what that rating was. Similar to the credit scores provided by Creditreform or
Dun & Bradstreet, the BoniCheck relies on past payment behavior, relative to
trade credit from utilities and suppliers. This information is complemented by
Buergel’s subjective assessment of the ventures’ future ability to fulfill credit
obligations, derived from information about the ventures’ order situation or
industry (Brown et al., 2012).
The distribution of good, fair, and bad credit scores is comparable within
both groups, exhibiting a total mean of 1.88, or a fair score on average. The
estimated marginal effect of credit ratings is significantly positive in all models.
A bad credit rating increases the probability of using crowdfunding by 31.2%
compared with ventures that have a fair rating.
4.1.2
Size
The decision to finance a venture is based on many factors. Larger ventures
can use economies of scale to reduce information asymmetries, but they also
have access with different sources of financing, because their risk exposure and
the scale of transaction costs differ. They often own more pledgeable collateral
and have more diverse cash flows. Small ventures instead are informationally
more opaque. Thus, size is an important choice factor when it comes to financ18
ing young ventures (Berger and Udell, 1998). Small ventures often struggle to
resolve informational asymmetries with investors and lenders at acceptable
costs, and they therefore are exposed to higher charges for smaller amounts of
capital. Transaction costs also influence funding methods. Small amounts often incur relatively high transaction costs, which is why some available sources
for certain kinds of ventures are not relevant (Titman and Wessels, 1988). For
example, the public issues of equity shares during an initial public offering
requires a scale that most small companies cannot reach in their early stages,
so small ventures are excluded from this type of financing (Cassar, 2004).
In summary, smaller ventures often face problems obtaining traditional
sources of outside financing, which could influence their use of crowdfunding. Empirical studies generally propose a positive link between venture size
and outside financing, leverage, and bank financing (Coleman, 2000; Cosh
et al., 2009). Therefore, we expect that smaller ventures are more likely to use
crowdfunding than large ones.
The mean size (log of total assets) of the sample ventures is 11.78. Ventures
that made no use of crowdfunding are larger in terms of total assets, with
a log of 12.35 (≈ EUR 230.000), than ventures that use of crowdfunding,
whose logged size was 10.99 (≈ EUR 60.000). We specify the log of assets to
measure size so that we can mitigate the influence of outliers in the skewed
size distribution. 2
The coefficient for size is negative and statistically significant in all models.
In line with the expected effect, the coefficient estimate indicates that smaller
2
Alternative treatments of the outliers, such as winsorizing, did not alter our results
qualitatively.
19
ventures are more likely to use crowdfunding; a greater size, in terms of logged
total assets, decreases the probability per unit change by 8%.
4.1.3
Tangibility
Another trait related to financing, particularly for young ventures, is the
structure of their assets (Cassar, 2004). In case a bankruptcy occurs, the financial loss for investors can be reduced if the assets are more tangible and
generic (Harris and Raviv, 1991; Titman and Wessels, 1988). Moreover, the
adverse selection and moral hazard costs should decrease when ventures pledge
assets as collateral or charges get fixed on the tangible assets. Tangible assets
increase liquidation value, so companies with a higher share of tangible assets
should gain access to traditional sources of finance more easily. The lower costs
of financing then tend to result in a higher degree of leverage in the capital
structure of these ventures. Empirical evidence suggests that banks base their
financing decision, to a certain degree on whether they can hedge the loan
with tangible assets (Berger and Udell, 1998; Storey, 1994). Considering the
substantial information asymmetries at the beginning of a venture’s life cycle and the information needed to forecast future development, investors have
relatively few ways to reduce their risk exposure, other than relationship banking. The asset structure, in terms of the share of tangible assets, often serves
as a screening tool for banks, such that it has significant effects on financing at the beginning of a venture (Cassar, 2004). Consistent with theoretical
predicitions, some authors suggest a positive relationship between the share
of tangible assets and leverage for large ventures, but research pertaining to
small ventures is rare, with a just a little evidence of a relationship between
the asset structure and the use of debt (e.g. Michaelas et al., 1999). Nevertheless, we expect that the lower the share of tangible assets of a venture, the
20
higher the likelihood of using crowdfunding.
To calculate the asset structure of each venture for every year since its
foundation, we divided the non-current assets by total assets, then take the
average of these values to define the variable (tangibility), ranging from ”0” to
”1”. For the entire sample, tangible assets constitute around 15% of the total
assets of the ventures, but among ventures that did not use crowdfunding, the
average tangible assets were greater 18% than it was for ventures that used
crowdfunding (10%).
The coefficients for the tangibility variable also were negative in all models
and significantly different from zero. Therefore, ventures with a lower share of
tangible assets appear to have a higher probability of using crowdfunding. A
1% decrease in the share of tangible assets increases the probability of using
crowdfunding by about 0.4%.
4.1.4
Characteristics of the Management Team and Venture
Financial ratios and external ratings alone cannot explain the financing
decisions of new ventures. Regarding young ventures in particular, many investors include the owner or management team in their assessment, because
their importance during the first years of operations cannot be underestimated (Cassar, 2004). For example, due to credit discrimination or the risk
aversion of some financiers, the gender composition of the management team
can influence the capital structure (Coleman, 2000). Arenius and Autio (2006)
provide evidence that female-owned businesses are often financed differently
than male-owned businesses. Other authors suggest that female-owned ventures have worse initial economic conditions, with a lower capital base (Verheul
and Thurik, 2001), and they face the problem of being less likely to obtain
21
external funding (Coleman, 2000). Furthermore, they usually use different
sources to finance their business than do male-owned ventures (Neider, 1987;
Lerner et al., 1997) and face particular difficulties applying for and securing
bank loans (Riding and Swift, 1990; Coleman, 2000; Anna et al., 2000). Ventures with mixed gender or purely female teams thus may be more likely to
use crowdfunding than ventures with a male management team.
The number of members in the management team also can affect the
chances of obtaining external capital. Chandler and Hanks (1998) show that
ventures founded and led by a team often are more successful than those
founded and led by single person. Beckman et al. (2007) find that the number of team members and the team composition have positive effects on the
likelihood of ventures attracting external financing. Therefore, we posit that
ventures with smaller management teams are more likely to use crowdfunding.
To control for management team characteristics, we add the number of
management team members (heads) and the gender composition of the management team. The latter is specified as an ordinal variable (gender ), with
”1” indicating a male-only team, ”2” a mixed team, and ”3” a female-only
team. Most of the management teams in the sample were purely male, as the
1.21 mean for the gender composition shows. Ventures that used crowdfunding
included slightly more women in their teams (1.38) than ventures that did not
use crowdfunding (1.1). With respect to the number of heads in the management team, the groups were comparable, with a total average of 1.59 persons,
though the ventures that did not use crowdfunding were slightly larger on
average.
The scholarship variable indicated whether the venture received support
from the Federal Ministry of Economics and Energy (“Bundesministerium für
22
Wirtschaft und Energie”, BMWi), in the form of an EXIST founder scholarship, a nationwide funding program, which supports innovative businesses that
started in universities and research institutions during their early phases, such
that it could be interpreted as a signal of quality. Approximately 20% of the
ventures in the sample received this kind of support, with similar distributions
across ventures that used and did not use crowdfunding.
Most financiers invest only within a close geographic scope (Gupta and
Sapienza, 1992), and rural areas are often characterized by worse access to
financing (Strotmann, 2006). Therefore, ventures from rural areas should exhibit a higher likelihood of using crowdfunding than ventures from urban areas.
The dichotomous variable (city) equals ”1” if the headquarters is located in
a city with more than 500,000 (urban) inhabitants and ”0” otherwise (rural).
Of all ventures, 73% are located in cities with more than 500,000 inhabitants,
though ventures that did not use crowdfunding were slightly more often located in urban areas (74%) than ventures that used crowdfunding (73%).
The hypothesis for the city variable predicted that ventures in rural areas
should have a higher likelihood of using crowdfunding, because they have
less access to finance. The coefficient for this variable was not statistically
significant though, the negative sign indicated that ventures in rural areas are
more likely to use crowdfunding.
In the last model in Table 6, the variables capturing the number of management heads and the location of the headquarters are both negative but not
significant. The existence of a scholarship seems to increase the probability of
using crowdfunding but is also not different from zero. The gender composition also has an important role, with a positive effect of using crowdfunding
when the management team is female. Compared with solely male teams, the
23
probability of using crowdfunding increases significantly by 28% when there
are male and female heads and by about 42% when the management team is
purely female.
4.1.5
Rating of Sophisticated Investors
A business plan is one of the most important steps to take when launching a
venture. In addition to providing economic efficiency, it exists mainly to raise
funds to start or expand a project. Mason and Harrison (1996) thus assert that
the business plan is the minimum requirement for any financing application,
because more than three-quarters of business angels base their investment
decision on this document. Different studies investigate the decision-making
process adopted by venture capital companies and suggest that the owner,
the business strategy, and financial issues are not the only determinants of
investment decisions (Zacharakis and Meyer, 1998; Hall and Hofer, 1993).
Many investors focus on product potential, industry-specific outlooks, and
growth opportunities. Sweeting (1991) shows that equity investors typically
spend less than 10 minutes on the first screening, and Hall and Hofer (1993)
indicate that they spend less than six minutes. Business angels typically devote
up to nine minutes to the screening process (Mason and Rogers, 1997). As an
emergent and therefore rather unusual tool for financing a venture due to
possible legal uncertainties, crowdfunding likely represents a second choice,
such that ventures likely tried to obtain funds through traditional sources of
capital first. Therefore, we expect that ventures that do not provoke detailed
investigation or consideration by sophisticated investors are more likely to use
crowdfunding.
The funding decision is often modeled as a stepwise process (Haines et al.,
24
2003), with three different phases: the initial screening, the detailed investigation, and the negotiation and deal closing. The information provided on
the ventures’ websites provides a way to imitate the screening process and
obtain ratings from different, sophisticated investors about the quality of the
ventures in the data sample, as well as whether they would move on to the
second step of the process, the detailed investigation, or would decide not to
pursue them after the first screening. To gather this measure, we presented all
the ventures to seven different equity investors from Germany, who indicated
if they would further investigate investing in each venture. To avoid bias, the
selected investors differed in their characteristics, such as deal volume, industry focus, type, and location. The variable (rating) is the sum of the single
ratings, which provide a dummy variable equal to ”1” for interesting followup investment opportunities and ”0” for ventures that the investors would
not take into consideration. The average rating was almost identical for both
groups. On average, about 50% of the investors would take a venture from the
sample into consideration for further investigation, and the difference between
the groups was small. This last explanatory variable controls for whether the
ventures used crowdfunding because they were not considered for detailed
investigation by sophisticated investors. Although the negative coefficient of
the rating variable indicated that ventures classified as non-qualified for further investigation by investors had a higher probability of using crowdfunding
than companies that were considered by more investors, the coefficient was
not significantly different from zero.
25
4.2
Bank characteristics
A possible concern in our analysis is that our results could be driven by
unobservable bank characteristics that may be correlated with the SoFFin
indicator and subsequent lending and risk taking. The SoFFin indicator could
therefore could merely confound unobserved traits with credit supply crunch
effects.
To mitigate this concern, we inclided bank-level control variables in a next
step, measured as the average over the period 2009-2014. They included the
same control variables we described previously, such that we considered various
proxies of financial health, measured according to the CAMEL supervisory
ratings system (i.e., capital, asset quality, management quality, and liquidity).
Table 7 reports the results of the baseline model with a stepwise integration
of CAMEL covariates.
– Table 7 around here –
The positive effect of crunched banks on the use of crowdsourced finance
indicated by support from the SoFFin, remained statistically and economically
significant even when we controlled directly for financial bank profiles. The
concern that the SoFFin indicator merely confoundsed unobserved traits as
credit supply shocks thus was invalidated by the intact, significant SoFFin
effect.
26
4.3
Alternative bank stress indicators
Some of the ventures were founded after the capital injections by the SoFFin, so that our results could driven by the ventures’ choice of a bank supported by the SoFFin, rather than a non-supported bank. Table 8 shows the
effect of the alternative bank stress indicators illustrated in Figure 2.
– Table 8 around here –
Column (1) replicates the baseline results with bank-specific controls. The alternative bank stress indicators in column (2) refer to the connection of one of
the local savings banks with a stressed Landesbank, as in Puri et al. (2011).
These authors show that local savings banks restricted loan supply when they
were connected to a Landesbanken with substantial subprime exposure. We
similarly include an indicator of whether a regional savings bank in our sample belonged to a stressed Landesbank. Although the marginal effect of the
Landesbanken variable was positive, indicating a higher probability of using
crowdfunding when the respective bank of a venture belonged to a stressed
Landesbank, the coefficient was very small and not significant.
Next, we included the results of the EU-wide bank stress test by the EBA,
published in November 2014, because it gauges information that was not available to ventures that might have selected banks on quality. First, we used the
EBA variable which indicates whether a bank was assessed directly or indirectly, and we split this sample into directly assessed banks and banks with
holdings in directly assessed banks in column (3). The direct assessment by the
EBA is only an indication of systemic importance, according to the regulator,
we separately specify in columns (4) and (5), whether a bank was assessed directly or merely connected to a holding company that was stress tested by the
27
EBA. The effect of being directly assessed by the EBA was generally similar
to the SoFFin indicator in that it was positive, but it could not be estimated
with sufficient precision to confirm statistical significance. Distinguishing between directly and and indirectly assessed banks shed further light on the EBA
assessment indicator: Whereas the effect of a direct assessement increased the
probability of using crowdfunding by about 10%, the indicator of indirectly
assessed banks was not significant.
In column (6), we specify a more direct measure of the health of the banks
tested (direct and indirectly). Financial institutions reported, during the comprehensive assessment in November 2014, whether they had a restructuring
plan in place before December 2013. The new ventures, sampled between 2011
and 2014, are unlikely to have had full knowledge of such restructuring initiatives when choosing whether to apply for crowdfunding, conditional on their
existing bank relationships. For the restructuring plan variable, the marginal
affirmed indeed that the probability of using crowdfunding increased by 17%.
Finally, in columns (7) through (10), we specify a range of interaction terms
of the direct assessment in the EBA test and the respective outcomes (i.e.,
CET1 ratio in in the adverse scenario 2016, the difference between the CET1
ration in the adverse scenario 2016 and the CET1 ratio starting in 2013,
and an binary indicator of whether the CET1 ratio in the adverse scenario
2016 was lower than 6%). These results corroborate prior findings that new
ventures with ties to banks that had restructuring plans in place and were
tested directly by EBA were more likely to tap crowdfunding.
In summary, for a range of alternative indicators of bank distress, and in
particular the existence of restructuring plans shared with the EBA, we found
results that were qualitatively similar to those we obtained with the SoFFin
28
indicator, albeit with some weaker statistical significance.
4.4
Using vs. succesfully completing crowdfunding
The previous analysis indicates that ventures are more likely to use crowdfunding when their bank is stressed. But applying for crowdfunding does not
automatically imply tthe successful completion of the funding request. Only
85% of the ventures in our sample were able to convince the crowd and collect
the minimum requested funding volume. Thus, the wisdom of the crowd may
be just as skilled as conventional intermediaries in selecting lemons out of the
pool of applicants.
To test this conjecture, we differentiated between ventures that applied for
crowdfunding and those that successfully obtained crowdfunding financing as
a function of stressed versus healthy bank relationships. With this information,
we provide more direct evidence of whether the wisdom of the (investor)crowd
can substitute for bank credit as a major funding source of new ventures if
banks are shocked.
– Table 9 around here –
In Table 9, we compare (1) the probability of applying for crowdfunding
with (2) the probability of successfully completing a crowdfunding request in
the full sample and (3) the probability of successfully completing a crowdfunding request among ventures that applied for crowdfunding. The relationship
with a stressed bank increased the probability of using crowdfunding in the
baseline model by 17%, and the same variable explained an increase of 22%
in the probability of successfully completing a crowdfunding request. The successful completion of a crowdfunding request among the 157 ventures only did
29
not depend on indicators of bank distress. Thus, credit supply shocks appear to
determine the choice to seek alternative funding forms, but do not necessarily
discriminate between projects that can or cannot convince the crowd.
5
Conclusion
Financing is a key component of entrepreneurial activities. By observing,
which ventures cooperated with banks that had to be bailed out by the German
government, we identify an effect of an exogenous credit supply shock on the
likelihood of using equity crowd funding. To this end, we manually collected
a unique data set that provided information about the financing decisions of
young ventures in Germany. Specifically, we used data from 357 young ventures
to test how certain characteristics, in terms of bank relationship, size, asset
structure, and other factors, affect the probability that the venture will use
crowdfunding.
Our results show that a relationship of a venture with a bailed out bank
increases the probability that a venture uses crowdfunding by 18%. This effect
is both economically and statistically significant. The analysis also shows that
bad credit scores increase the probability that a venture uses crowdfunding
by 31%. Supply-side restrictions move banks to handle their lending more
restrictively, and ventures that cannot demonstrate their creditworthiness are
not financed. This result suggests that among opaque new ventures, riskier
projects tend to tap equity crowdfunding instead of bank financing.
We also find that smaller ventures and ventures with fewer tangible assets are more likely to use crowdfunding. The small amounts obtained in a
crowdfunding offering makes this finding plausible. Larger ventures often need
30
greater volumes and have access to other or cheaper sources of capital, such
as initial public offerings. However, management team characteristics have no
statistically significant effect. Likewise, the rating of the venture’s quality by
experts, the location of the headquarters, the receive of a scholarship, and
the number of heads all showed no significant influence on a venture’s use of
crowdfunding. That is, the use of crowdfunding is not a question of management or other organizational factors. This result also supports the hypothesis
that quality differences are not crucial.
Perhaps the most important finding though is that ventures are more likely
to use crowdfunding when their bank is affected by a credit crunch. Equity
crowdfunding thus seems to be of particular importance for entrepreneurial
finance, as a critical source of capital in stressful times for banks.
31
Table 1
German Crowdfunding Market Overview
Year
2011
2012
2013
2014
Total
-
-
0.18
0.37
0.55
(2)
(4)
(6)
3.0
1.1
4.1
(1)
(1)
(2)
0.55
2.65
3.9
7.1
(6)
(15)
(9)
(30)
-
0.56
0.48
1.04
(5)
(6)
(11)
Platform
Bankless24
Bergfuerst
Companisto
Fundsters
Innovestment
Mashup Finance
Seedmatch
Others
Total
-
-
-
-
0.1
1.0
0.85
0.3
2.25
(2)
(13/8)
(11/4)
(7)
(33/12)
-
0.1
0.11
-
0.21
(1)
(1)
0.35
2.2
7.32
9.17
19.04
(4)
(22)
(22/1)
(20)
(68/1)
-
0.0
0.55
0.45
1.0
(1)
(11)
(7)
(19)
0.45
3.85
15.22
15.77
35.29
(6)
(43/8)
(68/5)
(54)
(171/13)
(2)
Notes: This table presents the volume raised in the German equity crowdfunding market with successful
campaigns, in millions of EUR, during the period 2011-2014. The number of (successful/unsuccessful) offerings
appear in brackets. Source: Own elicitation.
32
Table 2
Descriptive Statistics: Bank Characteristics by Stress Indicator
Mean
Capital
0.074
Asset Quality
0.002
Management
0.726
Earnings
0.036
Liquidity
0.191
Sec./Ear. Assets 0.273
Fees/Interest
0.307
Observations
Mean
Capital
0.072
Asset Quality
0.004
Management
0.762
Earnings
0.034
Liquidity
0.221
Sec./Ear. Assets 0.290
Fees/Interest
0.316
Observations
Capital
Asset Quality
Management
Earnings
Liquidity
Sec./Ear. Assets
Fees/Interest
Observations
Mean
0.078
0.006
0.782
0.043
0.221
0.284
0.336
Mean
Capital
0.076
Asset Quality
0.002
Management
0.723
Earnings
0.036
Liquidity
0.174
Sec./Ear. Assets 0.262
Fees/Interest
0.303
Observations
Mean
Capital
0.072
Asset Quality
0.005
Management
0.789
Earnings
0.036
Liquidity
0.249
Sec./Ear. Assets 0.306
Fees/Interest
0.335
Observations
Capital
Asset Quality
Management
Earnings
Liquidity
Sec./Ear. Assets
Fees/Interest
Observations
Mean
0.066
0.000
0.675
0.031
0.169
0.260
0.299
Soffin
yes
p95 Mean
SD
p5
p95
0.097 0.036 0.009 0.008 0.097
0.059 0.005 0.004 0.029 0.008
0.801 0.954 0.310 0.734 1.17
0.069 -0.110 0.125 0.199 0.022
0.571 0.268 0.225 0.109 0.428
0.502 0.418 0.160 0.305 0.531
0.472 0.237 0.397 0.044 0.517
2
Landesbanken
no
yes
SD
p5
p95 Mean
SD
p5
p95
0.047 0.095 0.095 0.075 0.021 0.041 0.098
0.018 0.007 0.009 -0.001 0.004 0.028 0.005
0.563 0.546 0.824 0.647 0.066 0.559 0.722
0.042 0.096 0.069 0.027 0.016 0.004 0.051
0.216 0.055 0.73 0.118 0.097 0.041 0.373
0.134 0.127 0.54 0.238 0.085 0.138 0.37
0.194 0.051 0.598 0.274 0.049 0.19 0.365
60
22
EBA
no
yes
SD
p5
p95 Mean
SD
p5
p95
0.054 0.043 0.097 0.067 0.022 0.029 0.096
0.021 0.052 0.008 -0.000 0.005 0.008 0.006
0.685 0.515 0.819 0.683 0.108 0.559 0.805
0.030 0.030 0.081 0.021 0.040 0.022 0.042
0.224 0.016 0.685 0.167 0.165 0.041 0.518
0.115 0.115 0.502 0.269 0.135 0.134 0.531
0.212 0.072 0.576 0.275 0.107 0.176 0.469
40
42
EBA direct
no
yes
SD
p5
p95 Mean
SD
p5
p95
0.041 0.042 0.070 0.032 0.014 0.003 0.007
0.016 0.008 0.008 0.004 0.003 0.001 0.008
0.500 0.539 0.797 0.838 0.166 0.734 1.17
0.025 0.009 0.019 -0.014 0.099 0.199 0.068
0.176 0.041 0.373 0.436 0.290
0.1 0.836
0.109 0.129 0.497 0.456 0.178 0.212 0.718
0.161 0.176 0.471 0.328 0.271 0.044 0.726
76
6
EBA indirect
no
yes
SD
p5
p95 Mean
SD
p5
p95
0.053 0.029 0.096 0.073 0.018 0.030 0.034
0.020 0.004 0.098 -0.001 0.005 0.011 0.006
0.641 0.539 0.841 0.658 0.070 0.558 0.793
0.047 0.005 0.046 0.027 0.015 0.004 0.051
0.242 0.066 0.836 0.122 0.072 0.04 0.302
0.136 0.135 0.531 0.237 0.098 0.129 0.437
0.217 0.055 0.679 0.266 0.048 0.19 0.347
46
36
Resturcturing Plan
no
yes
SD
p5
p95 Mean
SD
p5
p95
0.024 0.043 0.012 0.069 0.021 0.072 0.11
0.004 0.005 0.005 -0.001 0.006 0.010 0.009
0.082 0.559 0.797 0.693 0.133 0.559 0.989
0.019 0.004 0.068 0.010 0.053 0.113 0.022
0.186 0.043 0.518 0.164 0.143 0.034 0.525
0.142 0.134 0.479 0.278 0.130 0.132 0.54
0.110 0.218 0.469 0.248 0.099 0.008 0.411
22
20
no
SD
p5
0.041 0.033
0.016 0.009
0.488 0.546
0.026 0.006
0.197 0.097
0.123 0.131
0.165 0.159
80
Total
Difference
Mean
0.073
0.003
0.731
0.032
0.193
0.276
0.305
SD
p5
0.041 0.032
0.015 0.007
0.484 0.553
0.037 0.003
0.197 0.043
0.125 0.134
0.169 0.142
82
Total
p95
0.096
0.008
0.805
0.068
0.518
0.506
0.473
Mean
0.073
0.003
0.731
0.032
0.193
0.276
0.305
SD
p5
0.041 0.032
0.015 0.007
0.484 0.553
0.037 0.003
0.197 0.043
0.125 0.134
0.169 0.142
82
Total
p95
0.096
0.008
0.805
0.068
0.518
0.506
0.473
Mean
0.073
0.003
0.731
0.032
0.193
0.276
0.305
SD
p5
0.041 0.032
0.015 0.007
0.484 0.553
0.037 0.003
0.197 0.043
0.125 0.134
0.169 0.142
82
Total
p95
0.096
0.008
0.805
0.068
0.518
0.506
0.473
Mean
0.073
0.003
0.731
0.032
0.193
0.276
0.305
SD
p5
0.041 0.032
0.015 0.007
0.484 0.553
0.037 0.003
0.197 0.043
0.125 0.134
0.169 0.142
82
Total
p95
0.041
0.043∗∗
0.008
-0.001
0.805
-0.115
0.068
0.049
0.518
-0.262∗
0.506 -0.195 ∗∗∗
0.473
-0.026
Mean
0.073
0.003
0.731
0.032
0.193
0.276
0.305
SD
p5
0.041 0.032
0.015 0.007
0.484 0.553
0.037 0.003
0.197 0.043
0.125 0.134
0.169 0.142
82
Total
p95
0.096
0.008
0.805
0.068
0.518
0.506
0.473
SD
p5
0.041 0.032
0.015 0.007
0.484 0.553
0.037 0.003
0.197 0.043
0.125 0.134
0.169 0.142
42
p95
0.096
0.008
0.805
0.068
0.518
0.506
0.473
Mean
0.073
0.003
0.782
0.032
0.193
0.276
0.305
0.038
-0.003
-0.228
0.146
-0.077
-0.146
0.070
Equal
Variances
yes
yes
yes
no
yes
yes
no
Difference
-0.003
0.005∗
0.116
0.007
0.103∗∗∗
0.052∗∗
0.042∗
yes
no
yes
yes
no
no
no
Difference
0.011
0.006∗
0.098
0.022∗∗∗
0.055
0.015
0.062∗
yes
yes
yes
yes
yes
yes
yes
Difference
yes
yes
yes
no
no
yes
no
Difference
-0.001
0.006∗
0.131
0.009
0.127∗∗∗
0.069∗∗∗
0.070∗∗
yes
yes
yes
no
no
no
no
Difference
-0.003
0.001
-0.018
0.021∗
0.005
-0.018
0.051
yes
yes
yes
yes
yes
yes
yes
Notes: Descriptive statistics for the characteristics of the banks in the full sample (82 banks, 357 relationships), as well
as separately for the different stress indicators. For each variable, this table presents the mean, standard deviation, 5th
percentile, 95th percentile, and difference-in-means.
33
Table 3
Descriptive Statistics: EBA Stress Tests
EBA indirect
CET1 16
Restructuring plan
CET1 16 - CET1 13
CET1 16 < 8%
Observations
Mean
0.073
0.472
0.050
0.861
EBA direct
Total
SD
p5
p95 Mean SD
p5
p95 Mean SD
p5
p95
0.009 0.055 0.085 0.071 0.018 0.055 0.107 0.073 0.010 0.055 0.085
0.506
0
1
0.500 0.548
0
1
0.476 0.505
0
1
0.018 0.016 0.080 0.044 0.023 0.018 0.080 0.049 0.019 0.016 0.080
0.351
0
1
0.833 0.408
0
1
0.857 0.354
0
1
36
6
42
Notes: These descriptive statistics refer to the results of the EBA, EU-wide stress tests conducted in
October 2014 on the bank level. The statistics refer to all tested banks in the full sample (42 banks,
282 relationships) and separately for ventures with a direct relationship to a tested bank (6 banks,
207 relationships) and a indirect relationship to a tested bank (36 banks, 75 relationships). For each
variable, the table presents the mean, standard deviation, 5th percentile, and 95th percentile.
34
Table 4
Descriptive Statistics of Ventures, Crowdfunding Projects, and Associated Banks
Crowdfunding
SoFFin
Venture Characteristics
Credit
Size
Tangibility
Gender
City
Heads
Rating
Scholarship
35
Crowdfunding Characteristics
CF Min. Amount
CF Max. Amount
CF Realized Amount
CF Success
Number of CF Investors
Firm Valuation before CF
Bank Characteristics
Capital
Asset Quality
Management
Earnings
Liquidity
Sec./Ear. Assets
Fees/Interest
Observations
Yes
No
Yes
No
Mean
SD
Mean
SD
Difference
in Means
2.043
10.866
0.123
1.449
0.638
1.464
3.275
0.203
0.629
1.576
0.137
0.777
0.484
0.655
1.235
0.405
2.143
11.226
0.069
1.286
0.857
1.629
2.486
0.200
0.692
1.315
0.089
0.667
0.355
0.843
1.337
0.406
-0.099
-0.360
0.054∗∗
0.164
-0.219∗∗
-0.165
0.790∗
0.003
yes
yes
no
yes
no
yes
yes
yes
69,936
271,052
216,295
0.797
289.672
2,253,699
116,214
434,322
385,257
0.405
328.759
2,719,899
53,164
272,857
233,773
0.914
315.000
1,907,944
20,022
495,471
498,250
0.284
337.048
1,131,583
16,771
-1,805.0
-17,478.1
-0.117∗
-25.33
345,754
yes
yes
yes
no
yes
no
0.049
0.002
0.726
0.047
0.386
0.427
0.414
0.023
0.004
0.085
0.025
0.326
0.204
0.210
0.030
0.008
0.747
-0.027
0.418
0.525
0.501
0.002
0.001
0.074
0.030
0.054
0.038
0.095
0.019∗∗∗
-0.006∗∗∗
-0.021
0.074∗∗∗
-0.032
-0.097∗∗∗
-0.087∗∗∗
no
no
yes
yes
no
no
no
69
35
Equal
Variances
Yes
Mean
SD
No
Mean
SD
Difference
in Means
Equal
Variances
Diff-in-Diff
1.661
12.317
0.184
1.081
0.742
1.653
3.419
0.218
0.524
1.826
0.212
0.351
0.439
0.722
1.362
0.414
1.810
12.513
0.216
1.000
0.810
1.667
3.381
0.143
0.750
1.586
0.204
0.000
0.402
0.796
1.396
0.359
-0.148
-0.196
-0.033
0.081∗∗
-0.068
-0.013
0.038
0.075
no
yes
yes
no
yes
yes
yes
yes
-0.049
0.165
-0.087
-0.083
0.152
0.151
-0.751∗
0.072
0.051
0.004
0.772
0.044
0.427
0.414
0.425
0.040
0.012
0.387
0.021
0.340
0.213
0.218
0.029
0.008
0.734
-0.022
0.428
0.531
0.517
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.0216∗∗∗
-0.004
0.038
0.066∗∗∗
-0.001
-0.117∗∗∗
-0.093∗∗∗
no
yes
yes
no
no
no
no
0.003
0.001
0.058
-0.009
0.031
-0.020
-0.005
124
21
Notes: Descriptive statistics for the outcome of the equity crowdfunding offerings, characteristics of the ventures, and bank characteristics over the period 2011-2014 in Germany on the
venture level, separated by the SoFFin indicator. The sample includes all ventures with no missing values. For each variable, the table presents the mean, standard deviation, 5th percentile,
95th percentile, difference-in-means, and difference-in-differences. A offering is successful when the realized amount is larger than the minimum amount requested. Monetary variables are in
thousands of EUR.
Table 5
Descriptve Statistics: Stress Indicators
Crowdfunding
no
SoFFin
Landesbanken
EBA
EBA direct
EBA indirect
Restructuring plan
Obs.
200
200
200
200
200
153
Mean
0.145
0.100
0.765
0.550
0.215
0.307
Total
yes
SD p5 p95 Obs. Mean
SD p5 p95 Obs. Mean
SD p5 p95
0.353
0
1
157 0.369 0.484
0
1
357 0.244 0.430
0
1
157 0.115 0.320
0
1
357 0.106 0.309
0
1
0.301
0
1
0.425
0
1
157 0.822 0.384
0
1
357 0.790 0.384
0
1
157 0.618 0.488
0
1
357 0.580 0.494
0
1
0.499
0
1
0.412
0
1
157 0.204 0.404
0
1
357 0.210 0.408
0
1
0.463
0
1
129 0.566 0.498
0
1
282 0.426 0.495
0
1
Notes: These descriptive statistics reflects the characteristics of the stress indicators in the full sample.
Statistics are presented for all ventures in the full sample (357 ventures) and separately for ventures that
used crowdfunding (157 ventures) over the period 2011-2014 and ventures that not use crowdfunding (200
ventures). For each variable, the table presents the mean, standard deviation, 5th percentile, and 95th
percentile.
36
Table 6
Marginal Effects for the Use of Crowdfunding
(1)
SoFFin
∗∗∗
(2)
(3)
(5)
(7)
0.285
0.253
0.263
0.270
(0.052)
(0.052)
(0.048)
(0.049)
0.188
(0.059)
∗∗∗
0.189
(0.062)
0.175∗∗∗
(0.058)
0.137∗∗
0.090
0.084
(0.058)
(0.058)
(0.058)
0.487∗∗∗ 0.427∗∗∗ 0.416∗∗∗
(0.082)
(0.086)
(0.087)
0.048
(0.063)
0.334∗∗∗
(0.100)
0.095
(0.065)
0.381∗∗∗
(0.102)
0.046
(0.062)
0.312∗∗∗
(0.100)
0.323∗∗
0.337∗∗
(0.118)
(0.117)
∗∗∗
0.393
0.399∗∗∗
(0.079)
(0.078)
-0.050
(0.053)
-0.008
(0.032)
-0.017
(0.016)
0.068
(0.058)
0.334∗∗
(0.121)
0.401∗∗∗
(0.101)
-0.003
(0.059)
-0.011
(0.037)
-0.019
(0.018)
0.015
(0.064)
-0.086∗∗∗
(0.014)
0.218
(0.137)
0.457∗∗∗
(0.093)
-0.059
(0.061)
-0.025
(0.038)
-0.024
(0.019)
0.037
(0.066)
0.284∗∗
(0.126)
0.409∗∗∗
(0.103)
-0.014
(0.058)
-0.005
(0.036)
-0.023
(0.017)
0.016
(0.062)
-0.078∗∗∗
(0.014)
-0.004∗∗
(0.002)
249
0.280
0.426
0.837
Gender
mixed
female
City
Heads
Rating
Scholarship
Size
Tangibility
Observations
Pseudo R2
Nagelkerke’s Pseudo R2
AUC
(6)
∗∗∗
bad
∗∗∗
(4)
∗∗∗
Credit
fair
∗∗∗
357
0.049
0.087
0.612
357
0.107
0.182
0.705
357
0.157
0.260
0.749
357
0.164
0.270
0.763
249
0.260
0.401
0.828
-0.006∗∗∗
(0.002)
249
0.202
0.323
0.777
Notes: This table presents the average marginal effects from logit regression, where the dependent
variable is the use of crowdfunding. The sample consists of the 157 ventures that used crowdfunding
and 200 ventures that did not use crowdfunding. The variable definitions are provided in the
Appendix. Standard errors appear in brackets. Significance levels are as follows: *=10%, **=5%,
***=1%.
37
Table 7
Marginal Effects: Use of Crowdfunding with Bank Characteristics
(1)
(2)
(3)
(4)
0.172∗∗∗
(0.061)
0.224∗∗∗
(0.068)
0.175∗∗∗
(0.058)
0.046
(0.063)
0.311∗∗∗
(0.100)
0.039
(0.063)
0.314∗∗∗
(0.099)
0.046
(0.063)
0.311∗∗∗
(0.100)
(5)
(6)
(7)
(8)
0.207∗∗
(0.099)
0.174∗∗∗
(0.058)
0.169∗∗∗
(0.059)
0.174∗∗∗
(0.059)
0.279∗∗
(0.138)
0.046
(0.062)
0.310∗∗∗
(0.100)
0.045
(0.062)
0.313∗∗∗
(0.100)
0.046
(0.062)
0.306∗∗∗
(0.100)
0.046
(0.063)
0.312∗∗∗
(0.100)
0.027
(0.062)
0.294∗∗∗
(0.099)
Treatment Effect
SoFFin
venture characteristics
Credit
fair
bad
Gender
mixed
female
City
Heads
Rating
Scholarship
Size
Tangibility
0.283∗∗
0.290∗∗
0.283∗∗
0.282∗∗
0.290∗∗
0.278∗∗
0.283∗∗
0.310∗∗
(0.126)
(0.124)
(0.126)
(0.126)
(0.126)
(0.127)
(0.126)
(0.125)
0.408∗∗∗ 0.404∗∗∗ 0.408∗∗∗ 0.404∗∗∗ 0.408∗∗∗ 0.409∗∗∗ 0.410∗∗∗ 0.391∗∗∗
(0.103)
(0.103)
(0.103)
(0.104)
(0.104)
(0.102)
(0.102)
(0.105)
-0.017
0.003
-0.012
-0.018
-0.009
-0.024
-0.016
-0.023
(0.061)
(0.059)
(0.059)
(0.059)
(0.060)
(0.061)
(0.060)
(0.064)
-0.006
-0.006
-0.006
-0.006
-0.006
-0.005
-0.005
0.002
(0.036)
(0.036)
(0.036)
(0.036)
(0.036)
(0.036)
(0.036)
(0.036)
-0.023
-0.020
-0.023
-0.023
-0.022
-0.023
-0.023
-0.020
(0.017)
(0.017)
(0.017)
(0.017)
(0.017)
(0.017)
(0.017)
(0.017)
0.016
0.015
0.016
0.016
0.015
0.017
0.016
0.010
(0.062)
(0.062)
(0.062)
(0.062)
(0.063)
(0.062)
(0.062)
(0.061)
∗∗∗
∗∗∗
∗∗∗
∗∗∗
∗∗∗
∗∗∗
∗∗∗
-0.078
-0.079
-0.077
-0.078
-0.077
-0.078
-0.078
-0.081∗∗∗
(0.014)
(0.014)
(0.014)
(0.014)
(0.014)
(0.014)
(0.014)
(0.014)
-0.004∗∗
-0.004∗∗
-0.004∗∗
-0.004∗∗
-0.004∗∗
-0.004∗∗
-0.004∗∗
-0.004∗∗
(0.002)
(0.002)
(0.002)
(0.002)
(0.002)
(0.002)
(0.002)
(0.002)
Bank characteristics
Capital
-0.002
(0.010)
Asset Quality
-0.100
(0.073)
Management
-0.000
(0.002)
Earnings
0.004
(0.011)
Liquidity
-0.000
(0.001)
Sec/Ear. Assets
0.001
(0.001)
Fees/Interest
Observations
Pseudo R2
Nagelkerke’s Pseudo R2
AUC
249
0.280
0.426
0.837
249
0.286
0.434
0.838
249
0.280
0.426
0.836
249
0.281
0.427
0.837
249
0.280
0.426
0.837
249
0.281
0.427
0.838
0.000
(0.001)
249
0.280
0.426
0.837
0.011
(0.016)
-0.137∗
(0.079)
0.004
(0.003)
0.010
(0.014)
-0.003
(0.002)
0.007
(0.004)
-0.002
(0.003)
249
0.299
0.450
0.847
Notes: This table presents the average marginal effects from a logit regression, where the dependent variable is the
use of crowdfunding. Independent variables include venture characteristics and characteristics of the respective
banks as CAMEL covariates. The sample consists of the 157 ventures that used crowdfunding and 200 ventures
that did not use crowdfunding. The variable definitions are provided in the Appendix. Standard errors appear
in brackets. Significance levels are as follows: *=10%, **=5%, ***=1%.
38
Table 8
Marginal Effects with Alternative Treatment Effects
(1)
SoFFin
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
(10)
0.016
(0.203)
0.009
(0.205)
∗∗∗
0.187
(0.057)
Landesbanken
0.006
(0.081)
EBA
0.030
(0.064)
EBA direct
0.106∗
(0.058)
EBA indirect
0.176∗ 1.220∗
(0.094) (0.726)
-0.113
(0.069)
Restructuring plan
EBA direct X Restructuring plan
0.169∗∗∗ 0.197∗∗
(0.054) (0.107)
-0.045
(0.127)
CET1 16
EBA direct X CET1 16
0.167∗
(0.091)
-0.139
(0.097)
CET1 16 - CET1 13
EBA direct X CET1 16 - CET1 13
-0.061∗
(0.032)
0.029
(0.039)
CET1 16 < 8%
EBA indirect X CET1 16 < 8%
Control variables included?
Observations
Pseudo R2
Nagelkerke’s Pseudo R2
AUC
yes
249
0.270
0.414
0.832
yes
249
0.242
0.378
0.823
yes
249
0.243
0.379
0.822
yes
249
0.252
0.390
0.826
yes
249
0.250
0.388
0.825
yes
196
0.296
0.447
0.846
yes
196
0.314
0.468
0.849
yes
196
0.298
0.449
0.844
yes
196
0.304
0.456
0.847
-0.276∗
(0.151)
0.198
(0.207)
yes
196
0.295
0.446
0.843
Notes: This table presents the marginal effects from a logit regression, where the dependent variable is the use of crowdfunding. Independent variables include the results of the EBA EU-wide stress test, respective interactions, and a Landesbanken
dummy. Venture characteristics are included in each model but are not shown. The sample consists of the 157 ventures that
used crowdfunding and 200 ventures that did not use crowdfunding. The variable definitions are provided in the appendix.
Standard errors appear in brackets. Significance levels are as follows: *=10%, **=5%, ***=1%.
39
Table 9
Marginal Effects: Use and Outcome of Crowdfunding
SoFFin
Credit
fair
bad
Gender
mixed
female
City
Heads
Rating
Scholarship
Size
Tangibility
Observations
Pseudo R2
Nagelkerke’s Pseudo R2
AUC
(1)
Use of Crowdfunding
(full sample)
(2)
Successful Use of Crowdfunding
(full sample)
(3)
Successful Use of Crowdfunding
(only ventures that used crowdfunding)
0.175∗∗∗
(0.058)
0.216∗∗∗
(0.058)
0.077
(0.087)
0.046
(0.062)
0.312∗∗∗
(0.100)
0.135∗∗
(0.064)
0.203∗∗
(0.098)
0.224∗∗
(0.110)
0.152
(0.124)
0.284∗∗
(0.126)
0.409∗∗∗
(0.103)
-0.014
(0.058)
-0.005
(0.036)
-0.023
(0.017)
0.016
(0.062)
-0.078∗∗∗
(0.014)
-0.004∗∗
(0.002)
249
0.280
0.426
0.837
0.318∗∗
(0.134)
0.333∗∗∗
(0.109)
0.019
(0.061)
0.035
(0.038)
-0.017
(0.018)
0.050
(0.066)
-0.041∗∗∗
(0.015)
-0.004∗∗
(0.002)
249
0.193
0.305
0.786
0.000
0.027
(0.091)
0.033
(0.071)
0.178∗∗
(0.085)
0.015
(0.022)
0.135
(0.100)
0.083∗∗∗
(0.024)
-0.003
(0.003)
95
0.370
0.482
0.886
Notes: This table presents the average marginal effects from a logit regression where the dependent variable is the use of
crowdfunding for the full sample ( Column (1) ), the successful use of crowdfunding (realized amount > requested amount) for
the full sample (Column (2) ), or the successful use of crowdfunding for the sample of ventures that used crowdfunding (Column
(3) ). The variable definitions are provided in the appendix. Standard errors appear in brackets. Significance levels are as follows:
*=10%, **=5%, ***=1%.
40
Fig. 1. Sample of New Ventures that Apply for Crowdfunding or Not
Notes: This figure shows the sample of ventures that applied to one of the six largest equity crowdfunding
platforms in Germany for funds between 2011 and 2014. Out of157 applicants, 133 ventures successfully
completed their funding request, 23 applying ventures did not complete requests, and 200 ventures did not
apply at all. Some ventures applied multiple times for funding. The data on non-applicants is obtained from
the German Federal Association of Startups. The data about crowdfunding applicants were collected from
observing applicant data directly in the online platforms maintained by Bankless24, Berfuerst, Companisto,
Fundsters, Innovestment, Mashup Finance, Seedmatch, and others.
41
Fig. 2. Definitions of Stressed Banks
Notes: This figure presents the 82 banks in the sample listed in Table 10 that are connected to the 357 ventures shown in Figure 1. The link between ventures and banks is collected from the Creditreform database.
The base line identification defines stressed banks as those that received equity support from the SoFFin.
Next, we also define banks as stressed if they had, according to the comprehensive assessment by the European Banking Authority (EBA) of 2014, a restructuring plan in place since before 2013. We distinguish
between banks assessed directly by the EBA (6) and those that were connected to a bank holding company
that was assessed. Finally, we consider all those regional savings banks that were connected to a Landesbank
distressed, because their responsible bank holding company was exposed to the U.S. subprime market shock.
(Puri et al., 2011).
42
Table 10
Bank Overview
BvD ID
13046
13047
13124
13190
13192
13216
13263
13264
13296
13319
13326
13331
13366
13379
13380
13400
13418
13437
13444
13498
13570
13655
13724
13727
13732
13740
13742
13762
13803
13804
13839
13842
13858
13866
13869
13885
13894
13896
13912
13937
14008
14011
14037
14067
14090
14104
14123
14133
14166
14199
14469
14530
14654
15415
27737
29867
40293
40583
40867
41395
42705
42771
43024
43128
43289
43393
43617
43968
44034
44155
44562
45341
45375
45877
46123
46801
47101
47634
47699
47734
49769
49838
Bank Name
Bank fuer Sozialwirtschaft
ING-DiBa
Volksbank Potsdam
Commerzbank
Donner & Reuschel
Deutsche Bank
Frankfurter Sparkasse Sprendlingen
Frankfurter Volksbank
Heidelberger Volksbank eG
IKB Deutsche Industriebank AG
Koelner Bank
Sparkasse Dachau
Kreissparkasse Ahrweiler
Sparkasse Zollernalb
Sparkasse Bamberg
Sparkasse Dueren
Kreissparkasse Gross-Gerau
Kreissparkasse Koeln
Kreissparkasse Ludwigsburg
Kreissparkasse Waiblingen
Nassauische Sparkasse
Sparkasse Bochum
Sparkasse Karlsruhe Ettlingen
Sparkasse Koblenz
Sparkasse Landshut
Sparkasse Mainz
Sparkasse Markgraeflerland
Sparkasse Passau
Stadt- und Kreis-Sparkasse Darmstadt
Stadt- und Kreissparkasse Erlangen
Sparkasse Aachen
Stadtsparkasse Augsburg
Sparkasse Harburg-Buxtehude
Stadtsparkasse Duesseldorf
Verbundsparkasse Emsdetten Ochtrup
Sparkasse Hannover
Kreissparkasse Kaiserslautern
Kasseler Sparkasse
Stadtsparkasse Muenchen
Stadtsparkasse Schwerte
Volksbank Ludwigsburg eG
Volksbank Paderborn-Hoexter-Detmold
Sparkasse Hoexter
Volksbank Stuttgart
Sparkasse Muelheim an der Ruhr
Berliner Sparkasse
Herner Sparkasse
Postbank
Volksbank Mittelhessen
Sparkasse Leipzig
Ostsaechsische Sparkasse Dresden
Volksbank Karlsruhe
Deutsche Kontor Privatbank AG
Raiffeisenbank Gundelfingen
National Bank
Sparkasse KoelnBonn
Hamburger Sparkasse
GLS Gemeinschaftsbank
Sparkasse Westmuensterland
VR-Bank Rhein-Sieg
Volksbank Rhein-Nahe-Hunsrueck
Sparkasse Oder-Spree
Volksbank Neckartal
Volksbank Erft
Nordthueringer Volksbank
Stadtsparkasse Magdeburg
VR-Bank Starnberg-Herrsching-Landsberg
Volksbank St. Blasien
Sparkasse Maerkisch-Oderland
VR-Bank Passau
Sparkasse Bremen
Raiffeisenbank Heinsberg
Sparkasse Herford
Raiffeisenbank Parsberg-Velburg
Volksbank Welzheim
HypoVereinsbank
VR Bank Muenchen Land
Volksbank Brilon-Baeren-Salzkotten
Vereinigte Volksbank Maingau
LBBW
Sparkasse Schaumburg
Volksbank Sauerland
Bank-Venture
Observations
2
1
13
84
2
77
2
1
1
3
1
1
1
1
2
1
2
2
1
1
1
1
2
2
1
2
1
1
1
3
1
1
1
3
1
4
2
1
7
1
1
1
1
2
1
31
1
18
1
1
2
1
1
2
1
4
10
5
1
1
1
1
1
1
1
2
1
1
1
1
1
1
1
1
2
11
1
1
1
2
1
1
Total Bank Observations
82
(357)
27
(47)
1
(2)
9
(198)
45
(110)
thereof cooperative banks
thereof landesbanken
thereof private banks
thereof savings banks
Category
cooperative
private
cooperative
private
private
private
savings
cooperative
cooperative
private
cooperative
savings
savings
savings
savings
savings
savings
savings
savings
savings
savings
savings
savings
savings
savings
savings
savings
savings
savings
savings
savings
savings
savings
savings
savings
savings
savings
savings
savings
savings
cooperative
cooperative
savings
cooperative
savings
savings
savings
private
cooperative
savings
savings
cooperative
private
cooperative
private
savings
savings
cooperative
savings
cooperative
cooperative
savings
cooperative
cooperative
cooperative
savings
cooperative
cooperative
savings
cooperative
savings
cooperative
savings
cooperative
cooperative
private
cooperative
cooperative
cooperative
landesbank
savings
cooperative
SoFFin
Landesbank
X
EBA direct
EBA indirect
X
X
X
X
X
Bayern LB
X
X
X
X
X
X
X
X
X
X
X
Bayern LB
West LB
West LB
West LB
Bayern LB
X
X
X
X
X
X
X
X
X
X
X
Bayern LB
Bayern LB
West LB
Bayern LB
West LB
West LB
Bayern LB
West LB
X
X
X
West LB
X
West LB
X
X
West LB
X
X
Sachsen LB
Sachsen LB
West LB
X
X
West LB
X
X
X
West LB
X
X
X
2
(87)
22
(38)
6
(207)
36
(75)
Notes: These descriptive statistics detail the banks in the full sample. The number of bank-venture observations appear in brackets.
43
Table 11
Definition of Variables
Variable name
Source
Crowdfunding characteristics
Crowdfunding
Crowdfunding
CF Min. Amount
Crowdfunding
CF Max. Amount
Crowdfunding
CF Realized Amount
Crowdfunding
CF Success
Crowdfunding
platforms
platforms
platforms
platforms
platforms
Number of CF Investors
Crowdfunding platforms
Valuation of Venture be- Crowdfunding platforms
fore CF
Venture characteristics
Size
Tangibility
Heads
Gender
Bundesanzeiger
Bundesanzeiger
Creditreform
Creditreform
Credit
Rating
Buergel
City
Creditreform
Scholarship
BMWi
Treatments
SoFFin
BMFS
Landesbanken
Sparkassen-Verband
EBA
European Banking Authority
EBA direct
European Banking Authority
EBA indirect
European Banking Authority
CET1 16
European Banking Authority
CET1 16 - CET1 13
European Banking Authority
CET1 16 < 8%
European Banking Authority
Restructuring plan
European Banking Authority
Bank Characteristics
Capital
Bankscope
Asset quality
Bankscope
Management
Bankscope
Earnings
Bankscope
Liquidity
Bankscope
Sec./Ear. Assets
Bankscope
Fees/Interest
Bankscope
Description
Measurement
unit
Dummy variable equal to one if the venture used crowdfunding
Minimum amount of the respective crowdfunding offering
Maximum amount of the respective crowdfunding offering
Realized amount of the respective crowdfunding offering
Dummy variable equal to one if the venture used successfully
crowdfunding (realized amount > minimum amount)
Number of investors in the respective crowdfunding offering
Valuation of Venture before the crowdfunding offering, which
is done by the platform
binary
EUR
EUR
EUR
binary
Log of total assets as average since foundation
Percentage of tangible assets as average since foundation
Number of heads in the management team
Gender
composition
of
the
management
team
(male/mixed/female)
Credit rating of the venture (good/fair/bad)
Rating of seven sophisticated investors if they would further
investigate an investment for each venture (0/1)
Dummy variable equal to one if the location of headquarter of
the venture is based in a city with more than 500,000 inhabitants
Dummy variable equal to one if the venture received the EXIST
scholarship by the BMWi
Log of EUR
%
#
categorical
Dummy variable equal to one if the bank of the venture received funds from the SoFFin
Dummy variable equal to one if the bank of the venture is a
savings bank that owns holdings in one of the affected Landesbanken (Bayern LB, Sachsen LB, West LB)
Dummy variable equal to one if the bank of the venture is
directly or indirectly included in the 2014 EU-wide stress test
conducted by the European Banking Authority (EBA)
Dummy variable equal to one if the bank of the venture is
directly included in the 2014 EU-wide stress test conducted by
the European Banking Authority (EBA)
Dummy variable equal to one if the bank of the venture is
indirectly over holdings included in the 2014 EU-wide stress
test conducted by the European Banking Authority (EBA)
Fully loaded Common Equity Tier 1 (CET1) ratio in the adverse scenario 2016
Difference between the CET1 ratio starting 2013 and the fully
loaded CET1 ratio in the adverse scenario 2016
Dummy variable equal to one if the CET1 ratio of the tested
bank is lower than 8% in the adverse scenario 2016
Dummy variable equal to one if the bank of the venture had
an restructuring plan before 2013
binary
Proxy for capital adequacy of a venture’s bank measured as
the ratio of total equity to total assets
Proxy for asset quality of a venture’s bank measured as the
ratio of loan loss provisions to total gross loans
Proxy for managerial quality of a venture’s bank measured as
the ratio of total costs to total income
Proxy for earnings of a venture’s bank measured as the return
on average equity
Proxy for liquidity of a venture’s bank measured as liquid assets to deposits and short-term funding
Proxy for liquidity of a venture’s bank measured as securities
to total earning assets
Non-interest income divided by net interest income
%
44
#
EUR
categorical
categorical
binary
binary
binary
binary
binary
binary
%
%
binary
binary
%
%
%
%
%
%
References
Akerlof, G. (1970). The market for lemons: Quality uncertainty and the market
mechanism. Quarterly Journal of Economics. 84 (3), 488–500.
Anna, A. L., G. N. Chandler, E. Jansen, and N. P. Mero (2000). Women
business owners in traditional and non-traditional industries. Journal of
Business Venturing 15 (3), 279–303.
Arenius, P. and E. Autio (2006). Financing of small businesses: Are Mars and
Venus more alike than different? Venture Capital 8 (2), 93–107.
Beckman, C. M., M. D. Burton, and C. O’Reilly (2007). Early teams: The
impact of team demography on VC financing and going public. Journal of
Business Venturing 22 (2), 147–173.
Belleflamme, P., T. Lambert, and A. Schwienbacher (2013, September).
Crowdfunding: Tapping the right crowd. Journal of Business Venturing 29 (5), 585–609.
Berger, A. and G. Udell (2004). The institutional memory hypothesis and the
pro-cyclicality of bank lending behavior. Journal of Financial Intermediation 13, 458–495.
Berger, A. N. and G. F. Udell (1998). The economics of small business finance:
The roles of private equity and debt markets in the financial growth cycle.
Journal of Banking and Finance 22 (6-8), 613–673.
Block, J., P. Sandner, and G. De Vries (2010). Venture capital and the financial
crisis: An empirical study across industries and countries. In D. Cumming
(Ed.), The Oxford Handbook of Venture Capital, Chapter 3. Oxford: Oxford
University Press.
Bradford, C. S. (2012). Crowdfunding and the federal securities law. Columbia
Business Law Review 2012 (1), 1–150.
Brown, M., H. Degryse, D. Höwer, and M. F. Penas (2012). How do banks
screen innovative firms? Evidence from start-up panel data. ZEW Discussion Papers No. 12-032, 1–37.
Cassar, G. (2004). The financing of business start-ups. Journal of Business
Venturing 19 (2), 261–283.
Chandler, G. N. and S. H. Hanks (1998). An examination of the substitutability of founders human and financial capital in emerging business ventures.
Journal of Business Venturing 13 (5), 353–369.
Coleman, S. (2000). Access to capital and terms of credit: A comparison
of men- and women-owned small businesses. Journal of Small Business
Management 38 (3), 37.
Cosh, A., D. Cumming, and A. Hughes (2009). Outside entrepreneurial capital. Economic Journal 119 (1), 494–533.
De Buysere, K., O. Gajda, R. Kleverlaan, and D. Marom (2014). A framework for european crowdfunding. Technical report, European Crowdfunding
Network.
de Meza, D. and D. C. Webb (1987). Too much investment: A problem of
asymmetric information. Quarterly Journal of Economics 102 (2), 281–92.
45
Dell’Ariccia, G. and R. Marquez (2004). Information and bank credit allocation. Journal of Financial Economics 72, 185–214.
Diamond, D. W. (1984). Financial intermediation and delegated monitoring.
Review of Economic Studies 51(3), 393–414.
Gompers, P. and J. Lerner (2001). The venture capital revolution. Journal of
Economic Perspectives 15 (2), 145–168.
Gorman, M. and W. Sahlman (1989). What do venture capitalists do? Journal
of Business Venturing 4 (4), 231–248.
Gupta, A. K. and H. J. Sapienza (1992). Determinants of venture capital
firms’ preferences regarding the industry diversity and geographic scope of
their investments. Journal of Business Venturing 7 (5), 347–362.
Haines, G. H., J. J. Madill, and A. L. Riding (2003). Informal investment in
Canada: Financing small business growth. Journal of Small Business and
Entrepreneurship 16 (3-4), 13–40.
Hall, J. and C. W. Hofer (1993). Venture capitalists’ decision criteria in new
venture evaluation. Journal of Business Venturing 8 (1), 25–42.
Harris, M. and A. Raviv (1991). The theory of capital structure. Journal of
Finance 46 (1), 297–355.
Hemer, J. (2011). A snapshot on crowdfunding. Working Paper Series: Firms
and Region - Fraunhofer ISI No. 2.
Hempell, H. S. and C. Kok (2010, November). The impact of supply constraints on bank lending in the euro area - crisis induced crunching? European Central Bank Working Paper Series No. 1262.
Hornuf, L. and A. Schwienbacher (2014). Crowdinvesting - Angel investing
for the masses? In Handbook of Research on Venture Capital: Volume 3.
Business Angels, pp. forthcoming. Cheltenham: Edward Elgar Publishing
Ltd.
Hosmer, D. W. and S. Lemeshow (2012). Applied Logistic Regression (3 ed.).
Hoboken, NJ: Wiley.
Jensen, M. C. and W. H. Meckling (1976). Theory of the firm: Managerial
behavior, agency costs and ownership structure. Journal of Financial Economics 3 (4), 305–360.
Jiménez, G., S. Ongena, J. Saurina, and J.-L. Peydró (2012). Credit supply
and monetary policy: Identifying the bank balance-sheet channel with loan
applications. American Economic Review 102 (5), 2301–2326.
Jiménez, G., S. Ongena, J. Saurina, and J.-L. Peydró (2014). Hazardous times
for monetary policy: What do twenty-three million bank loans say about the
effects of monetary policy on credit risk? Econometrica 82 (2), 463–505.
Lerner, M., C. Brush, and R. Hisrich (1997). Israeli women entrepreneurs: An
examination of factors affecting performance. Journal of Business Venturing 12 (4), 315–339.
Mason, C. and R. Harrison (1996). Why ’business angels’ say no: A case study
of opportunities rejected by an informal investor syndicate. International
Small Business Journal 14 (2), 35–51.
Mason, C. and A. Rogers (1997). What do investors look for in a business
46
plan? An exploratory analysis. In D. Deakins, P. Jennongs, and C. Mason
(Eds.), Entrepreneurship in the 1990s, pp. 29–46. London: Paul Chapman
Publishing.
Michaelas, N., F. Chittenden, and P. Poutziouris (1999). Financial policy and
capital structure choice in U.K. SMEs: Empirical evidence from company
panel data. Small Business Economics 12 (2), 113–130.
Mollick, E. (2014). The dynamics of crowdfunding: An exploratory study.
Journal of Business Venturing 29 (1), 1–16.
Mollick, E. R. (2013). Swept away by the crowd? Crowdfunding, venture capital, and the selection of entrepreneurs. SSRN Electronic Journal (Available
at SSRN, September 2014).
Myers, S. C. and N. S. Majluf (1984). Corporate financing and investment
decisions when firms have information that investors do not have. Journal
of Financial Economics 13 (2), 187–221.
Neider, L. (1987). A preliminary investigation of female entrepreneurs in
Florida. Journal of Small Business Management 25, 22–29.
Petersen, M. A. and R. G. Rajan (1994). The benefits of lending relationships:
Evidence from small business data. Journal of Finance 49 (1), 3–37.
Petersen, M. A. and R. G. Rajan (2002). Does distance still matter: The
information revolution in small business lending. Journal of Finance 57,
2533–2570.
Popov, A. and G. F. Udell (2012). Cross-border banking, credit access, and
the financial crisis. Journal of International Economics 87 (1), 147–161.
Puri, M., J. Rocholl, and S. Steffen (2011). Global retail lending in the aftermath of the US financial crisis: Distinguishing between supply and demand
effects. Journal of Financial Economics 100, 556–578.
Rajan, R. G. (1992). Insiders and outsiders: The choice between informed and
arm’s-length debt. Journal of Finance 47, 1367–1400.
Riding, A. L. and C. S. Swift (1990). Women business owners and terms
of credit: Some empirical findings of the Canadian experience. Journal of
Business Venturing 5 (5), 327–340.
Robb, A. M. and D. T. Robinson (2014). The capital structure decisions of
new firms. Review of Financial Studies 27 (1), 153–179.
Schwienbacher, A. (2013). Financing the business. In T. Baker and F. Welter
(Eds.), The Routledge Companion to Entrepreneurship. London: Routledge.
Schwienbacher, A. and B. Larralde (2010). Crowdfunding of small entrepreneurial ventures. In Handbook of Entrepreneurial Finance. Oxford
University Press.
Stiglitz, J. E. and A. Weiss (1981). Credit rationing in markets with imperfect
information. American Economic Review 71 (3), 393–410.
Storey, D. J. (1994). New firm growth and bank financing. Small Business
Economics 6 (2), 139–150.
Strotmann, H. (2006). Entrepreneurial survival. Small Business Economics 28 (1), 87–104.
Sweeting, R. (1991). UK venture capital funds and the funding of new
47
technology-based businesses: Process and relationships. Journal of Management Studies 6 (November), 601–622.
Titman, S. and R. Wessels (1988). The determinants of capital structure
choice. Journal of Finance 43 (1), 1–19.
Uchida, H., G. F. Udell, and N. Yamori (2012). Loan officers and relationship
lending to SMEs. Journal of Financial Intermediation 21 (1), 97–112.
Verheul, I. and R. Thurik (2001). Start-up capital: ”Does gender matter?”.
Small Business Economics 16 (4), 329–346.
Zacharakis, A. L. and G. Meyer (1998). A lack of insight: Do venture capitalists really understand their own decision process? Journal of Business
Venturing 13 (1), 57–76.
48